Belzutifan’s adjuvant kidney cancer benefit looks front-loaded, not evenly sustained
A reconstructed individual patient data analysis of phase 3 adjuvant renal cell carcinoma trials suggests the disease-free survival advantage seen with belzutifan plus pembrolizumab may be strongest while treatment is ongoing and less stable after the planned 1-year course ends. The signal is clinically important because adjuvant therapy is often judged by a single hazard ratio, yet cancer recurrence risk does not move at a uniform pace across follow-up. Here, the treatment effect appears to change as time passes, raising the possibility that early separation on the Kaplan-Meier curves may overstate how durable the benefit remains once the drug is stopped.
The analysis centered on LITESPARK-022, where central review showed a 24-month restricted mean survival time difference of 1.34 months, with a 95% confidence interval of 0.73 to 1.95. In plain clinical terms, that translates to about 6 weeks of additional average time without recurrence over 2 years. The gain was still present after month 12, but the incremental advantage slowed, with a post-12-month ΔRMST increase of 0.74 months per year, confidence interval 0.32 to 1.19. The pattern suggests the effect was not simply accumulating at a constant pace across the whole follow-up window.
The hazard ratio changed sharply once planned therapy ended
The most striking finding was the contrast between the treatment period and the post-treatment period. In LITESPARK-022, the hazard ratio for disease-free survival was 0.55 before month 12, with a 95% confidence interval of 0.43 to 0.71. After month 12, the hazard ratio shifted to 1.16, with a 95% confidence interval of 0.87 to 1.54, and the post-to-pre ratio reached 2.10. Investigator assessment followed the same general pattern, with a similar attenuation after treatment ended.
That shift matters because adjuvant therapy in renal cell carcinoma is meant to suppress microscopic residual disease at the point when the disease burden is low. A drug such as belzutifan, which acts on the hypoxia-inducible factor pathway, may blunt the biological machinery that allows residual clones to persist or expand after nephrectomy. If the benefit weakens once exposure stops, the most plausible explanation is not that the early signal was false, but that the suppression is time-limited and may depend on continued pharmacologic pressure.
The paper compared this pattern with other adjuvant renal cell carcinoma trials. In KEYNOTE-564 at its first interim analysis, the ratio was 2.02, and benefit persisted. In SORCE and PROTECT, the ratios were 2.09 and 2.59, and the apparent benefit faded. Those comparisons place LITESPARK-022 in a broader trial pattern: early benefit does not guarantee the same effect after protocol-defined treatment has ended.
Absolute disease-free survival gaps moved as follow-up extended
A post hoc analysis from months 12 to 30 sharpened the uncertainty. With central review, the disease-free survival difference changed by minus 3.7 percentage points, with a confidence interval from minus 8.1 to 0.8. Under investigator assessment, the corresponding change was 0.6 percentage points, with a confidence interval from minus 4.0 to 5.0. Those two estimates point in different directions, and the spread between them is a reminder that endpoint assessment can influence the apparent durability of adjuvant benefit.
This is not a trivial statistical wrinkle. In kidney cancer trials, recurrence events are often sparse early on and then cluster later as surveillance continues. A treatment effect that is visible during active dosing can compress once the drug is removed, especially if the biology of residual disease depends on the same pathway the drug was suppressing. The analysis therefore argues against treating a single hazard ratio as a durable summary of the entire adjuvant experience.
One sentence captures the central tension. The benefit was strongest before month 12.
Reconstructed data sharpen the signal, but they also define the limits
The investigators reconstructed individual patient data from published Kaplan-Meier curves from ten phase 3 adjuvant renal cell carcinoma trials. They estimated restricted mean survival time differences, hazard ratios before and after the planned treatment period, and disease-free survival differences within each trial. Because the data were reconstructed from published aggregate curves rather than sourced from original patient-level files, the analysis depends on the quality of the published plots, the completeness of numbers at risk, and the assumptions required to rebuild the underlying time-to-event data.
Follow-up was also short. That limitation is particularly relevant for adjuvant renal cell carcinoma, where late recurrences can occur well after the treatment window. A trial can show a clean early advantage and still leave open the question of whether the gap remains clinically meaningful once observation lengthens. Here, the data suggest the effect may be waning or at least becoming less stable after the medication is stopped, but they do not prove that the drug has no lasting role.
The patient-facing summary attached to the preprint states that adding belzutifan to pembrolizumab increased the average time without cancer returning by about 6 weeks over 2 years, while also causing more serious side effects. Those safety details are not expanded in the abstract, yet they sit at the center of any adjuvant decision: a modest average gain must be weighed against treatment burden when the patient is otherwise rendered free of visible disease after surgery.
Single-HR projections look too blunt for this setting
The clearest practical lesson is methodological. A single hazard ratio can flatten a time-varying effect and make an adjuvant regimen look more durable than it really is, or less durable than it is, depending on when the events accrue. Restricted mean survival time helped show the size of the benefit over a fixed horizon, while the pre- and post-treatment hazard ratios exposed a change in treatment effect over time. In a disease where recurrence risk shifts with surveillance duration and microscopic disease biology, those distinctions are not academic.
For clinicians, the analysis argues for caution when interpreting early DFS separation as proof of long-term persistence after treatment ends. For trialists, it supports reporting time-stratified estimates alongside standard overall hazard ratios, especially in studies where the intervention is stopped long before the final disease-free survival readout. The uncertainty is not about whether belzutifan has activity; it is about how long that activity lasts once adjuvant dosing stops, and whether the absolute DFS advantage can be expected to remain intact beyond the planned treatment period.